当前位置:
X-MOL 学术
›
Energy Build.
›
论文详情
Our official English website, www.x-mol.net, welcomes your
feedback! (Note: you will need to create a separate account there.)
A novel building sampling approach leveraging data mining and stratified sampling theory for energy optimization
Energy and Buildings ( IF 6.6 ) Pub Date : 2025-01-25 , DOI: 10.1016/j.enbuild.2025.115366
Zhijian Fang, Lei Lei, Run Zheng
Energy and Buildings ( IF 6.6 ) Pub Date : 2025-01-25 , DOI: 10.1016/j.enbuild.2025.115366
Zhijian Fang, Lei Lei, Run Zheng
Buildings play a significant role in global energy consumption, and effective sampling methods are essential for developing targeted, impactful energy conservation measures. This study proposes a novel building sampling method integrating data mining technology and stratified sampling theory to enhance the representativeness and accuracy of building energy consumption analysis. Utilizing an optimized Gaussian mixture model (GMM) and rough set (RS) theory, this research establishes a refined stratification model that effectively categorizes buildings based on energy consumption data. The research employs a dual-layered approach: initially, building energy data are clustered using GMM, which handles multiple variables to identify homogenous building groups. Subsequently, RS theory is applied to minimize redundancy by narrowing down influential factors, thus refining the stratification process. One of the key challenges in building energy analysis is determining the optimal sample size that balances accuracy with practical feasibility. This study tackles this by introducing an iterative approximation method that adjusts the sample size based on the coefficient of variation (CV) across different strata until the desired accuracy is achieved. The effectiveness of this method was validated using a dataset comprising 200 public buildings. Results demonstrate that the proposed method not only achieves precise stratification but also significantly reduces the necessary sample size for reliable energy consumption analysis. This reduction enhances both the representativeness of sampling results and the efficiency of the data collection process, effectively lowering time and economic costs. Moreover, it mitigates biases introduced by subjective factor selection, ensuring a more accurate determination of the optimal sample size.
中文翻译:
一种利用数据挖掘和分层采样理论进行能源优化的新型建筑采样方法
建筑物在全球能源消耗中发挥着重要作用,有效的采样方法对于制定有针对性的、有影响力的节能措施至关重要。该文提出了一种融合数据挖掘技术和分层抽样理论的新型建筑抽样方法,以增强建筑能耗分析的代表性和准确性。利用优化的高斯混合模型 (GMM) 和粗糙集 (RS) 理论,本研究建立了一个精细的分层模型,该模型根据能耗数据对建筑物进行了有效的分类。该研究采用了双层方法:最初,建筑能源数据使用 GMM 进行聚类,GMM 处理多个变量以识别同质建筑组。随后,应用 RS 理论通过缩小影响因素来最小化冗余,从而完善分层过程。建筑能量分析的主要挑战之一是确定平衡准确性和实际可行性的最佳样本量。本研究通过引入一种迭代近似方法来解决这个问题,该方法根据不同层的变异系数 (CV) 调整样本量,直到达到所需的精度。使用包含 200 座公共建筑的数据集验证了这种方法的有效性。结果表明,所提出的方法不仅实现了精确的分层,而且显著减少了可靠能耗分析所需的样本量。这种减少增强了采样结果的代表性和数据收集过程的效率,从而有效地降低了时间和经济成本。 此外,它还减轻了主观因子选择引入的偏差,确保更准确地确定最佳样本量。
更新日期:2025-01-25
中文翻译:

一种利用数据挖掘和分层采样理论进行能源优化的新型建筑采样方法
建筑物在全球能源消耗中发挥着重要作用,有效的采样方法对于制定有针对性的、有影响力的节能措施至关重要。该文提出了一种融合数据挖掘技术和分层抽样理论的新型建筑抽样方法,以增强建筑能耗分析的代表性和准确性。利用优化的高斯混合模型 (GMM) 和粗糙集 (RS) 理论,本研究建立了一个精细的分层模型,该模型根据能耗数据对建筑物进行了有效的分类。该研究采用了双层方法:最初,建筑能源数据使用 GMM 进行聚类,GMM 处理多个变量以识别同质建筑组。随后,应用 RS 理论通过缩小影响因素来最小化冗余,从而完善分层过程。建筑能量分析的主要挑战之一是确定平衡准确性和实际可行性的最佳样本量。本研究通过引入一种迭代近似方法来解决这个问题,该方法根据不同层的变异系数 (CV) 调整样本量,直到达到所需的精度。使用包含 200 座公共建筑的数据集验证了这种方法的有效性。结果表明,所提出的方法不仅实现了精确的分层,而且显著减少了可靠能耗分析所需的样本量。这种减少增强了采样结果的代表性和数据收集过程的效率,从而有效地降低了时间和经济成本。 此外,它还减轻了主观因子选择引入的偏差,确保更准确地确定最佳样本量。